Receive our newsletter - data, analysis and deep insights on FDI delivered to you

GDP is flawed as a measurement, so why is it still in use?

Using GDP to measure a country's progress stemmed from the Great Depression, and it is now widely accepted as being out of date, yet it is proving a hard habit to break.

Alternatives to GDP have been available for decades, yet the measure continues to be favoured by economists, politicians and the press. (Photo by Sean K/Shutterstock)

Gross domestic product, or GDP, is everywhere. No other statistics are reported on as excitedly by newspapers, watched as nervously by investors or exploited as ruthlessly by political opponents. None are collected so often, to such uniform standards and by so many governments, or used in such a wide array of policy instruments.

Yet, we are constantly told that GDP is inadequate, outdated and damaging. Alternatives have proliferated for decades, from Bhutan’s Gross National Happiness index, first mooted in 1972, to the UN’s Human Development Indicators, launched in 1990.

More recently, former French President Nicolas Sarkozy’s landmark 2010 ‘inquiry into happiness’, led by three world-famous economists, has failed to dislodge the dominance of GDP, while ex-UK Prime Minister David Cameron’s well-publicised ‘happiness agenda’, launched that same year, has faded into near-obscurity.

White papers from our partners

The rise of GDP

In the 1930s, the Great Depression prompted a frantic search among US economists and politicians to gauge just how wounded the economy was, and how much fiscal stimulation was necessary to restore it to health.

“Societies before the Great Depression were less focused on governing by statistics,” says the Dutch economist Rutger Hoekstra, author of Replacing GDP by 2030 and coordinator of the UN University’s WiSE Transformation Initiative. 

“Older attempts at calculating national accounts date back to the 18th century, but these were basically just one-off projects commissioned, for instance, by governments to see if they could afford to go to war.”

By the 1930s, however, the need for macroeconomic policymaking had become evident. It was in this context that the US Congress ordered economists at the Department of Commerce, working under Soviet émigré Simon Kuznets, to come up with a unified measure of national production.

What emerged from this work was GDP, a statistic designed to measure the monetary value of all final goods and services produced by an economy’s residents. This can be measured by adding up companies’ net revenues, consumers’ spending or residents’ incomes. (Theoretically, these methods should all produce at the same number – everyone’s spending is someone else’s income, after all.)

The statistic would go global after the Second World War. With the US pumping huge sums of cash into European countries through the Marshall Plan, the OECD was tasked with collating and disseminating comparable statistics on the effectiveness of the programme, for which they adopted GDP.

By that time, however, Kuznets had lost control of the project. The Department of Commerce had increasingly come under the influence of the British economist Richard Stone. 

While Kuznets was concerned about ensuring the statistic provided a decent representation of social welfare, Stone was more concerned with its usefulness as a policymaking tool. This meant measuring what could be accurately measured and what could be shaped by the policy levers available at the time.

The father of GDP was alert to its flaws

Nonetheless, GDP has come to play a major role in public perception and political discourse as an indicator of general prosperity. 

One problem that Kuznets had identified with this misuse of the measure was the inclusion in GDP of activities that do nothing to increase social welfare.

Implicit in GDP is the idea that any activity that turns a smaller amount of money into a bigger amount of money is productive. The greater the difference between the two, or ‘value added’, the more productive the activity is. GDP therefore includes many of what economists call “defensive activities” – activities that are not a source of value in themselves but are, unfortunately, necessary to enable value-creation elsewhere.

Prisons, militaries and hospitals, for instance, produce no value in themselves – at best, they enable the production of value elsewhere in society. Yet they do produce ‘value added’.

As a result, an American can contribute more to his or her country’s GDP by being imprisoned than by getting a job: the average salary of a Walmart inventory associate is about $20,000, much less than the $31,000 it costs to keep someone in prison for a year.

The effect of counting defensive activities in GDP can be significant for some countries. Oman spent $6.7bn on its military in 2020, according to the Stockholm International Peace Research Institute – close to 10% of its GDP. Russia’s GDP increased by $30bn between 2018 and 2019, partly due to $5.2bn in arms sales.

Kuznets also criticised the inclusion of certain activities that ought to be classified as intermediate inputs into production, such as commuting. Commuting is an expense incurred by workers as a cost of production (it is hardly something they undertake for pleasure), yet is counted in GDP as a consumption good.

The same criticism could be levied at the inclusion of financial and business services, which constitute an increasingly large proportion of many countries’ GDP. Financial institutions may enable a more efficient allocation of capital, but their activities are not directly a source of utility. The absurdity of this is evident in the fact that a prospective homeowner will raise their country’s GDP more by taking out an expensive mortgage than by buying their house outright.

Is social welfare excluded in GDP measurements?

Kuznets also expressed serious concerns about the exclusion of certain activities that evidently do add to social welfare. GDP is the aggregate of domestic value added, where value added is defined as the monetary value of final goods and services sold, minus the monetary value of their input. By definition, therefore, GDP excludes any good or service that is not sold for money.

An obvious casualty of this method of accounting is public services provided free at the point of use. A cardiac surgery in the UK, provided free of charge by the National Health Service, produces no value added – money is used to purchase inputs, but no money is exchanged for the output. By contrast, the same surgery in the US might fetch a healthy price, raising the US GDP substantially.

One way statisticians get around this problem is by pricing public services by the cost of their inputs. While obviously better than excluding them altogether, doing so inevitably undervalues publicly provided services over privatised ones.

Other activities, however, do not even involve monetary transactions at the input stage. Unpaid domestic labour, subsistence farming and non-market leisure activities are all invisible in most GDP statistics.

One study used surveys to gauge individuals’ time spent each day on unpaid domestic work, and then multiplied this by the average wage of a house worker to estimate the market value of this work. They found that unpaid household work is equivalent to about one-third of GDP in France, Finland and the US.

It was the decision to omit unpaid domestic work from GDP that eventually prompted Kuznets to quit his post at the Department of Commerce.

GDP was designed for a different era

More recent critics of GDP have taken issue with its failure to account for capital depreciation or resource depletion. As a ‘gross’ measure, GDP does not take into account the depreciation of a country’s capital goods, such as machinery and vital infrastructure, meaning that spending on maintenance and repairs is included in GDP.

For example, were the British government to dismantle the Tower of London brick by brick and sell the debris for scrap, the Office for National Statistics would register this as an increase in GDP.

The absence of capital depreciation from GDP statistics can make comparisons between countries and industries problematic, rewarding the adoption of more rapidly depreciating capital goods.

Computer hardware and software tends to depreciate much faster than previous generations of capital goods. As a result, digitisation provides an artificial boost to a country’s GDP, since more must be spent on replacing outdated goods.

Equally, GDP does not take into account actions that increase the value of a country’s assets, except insofar as these actions constitute a form of consumption. When Western volunteers travel to low-income countries to mend wells and orphanages for free, the only increase to their host country’s GDP comes from the bottled water and sandwiches they buy during their breaks.

Net versus gross national income

The significance of such omissions can be seen in the difference between a country’s net national income and its gross national income, with the former attempting to take into account the consumption of fixed capital.

When the value of depreciated capital, such as worn-out machines and outdated software, is subtracted from Ireland’s GDP, it falls by 32%. The figure is higher for countries that make greater use of digital equipment and software, since these become defunct much faster than traditional capital goods, such as heavy machinery.

Even measures such as net national income (NDI) typically fail to take account of the depletion of a country’s natural resources, however. For every Angolan diamond mined and sold, the number of diamonds remaining is diminished. From the perspective of GDP or NDI, however, the diamonds may as well be produced in factories. Similarly, whether a timber company simply cuts its way through the Amazon or nurtures a sustainable plantation has no bearing on its contribution to GDP.

The fact that GDP measures production, rather than income, has also become more problematic in recent decades, as globalisation has caused consumption to become ever more spatially separated from production.

Until 1993, the US used gross national product (GNP) rather than GDP, the difference being that GNP includes the value added of US citizens abroad.

At that time, US companies operating abroad repatriated just $29bn in profits per year – less than what they repatriate every month today. The profits repatriated in 2020 alone ($342bn) are larger than the sum repatriated in the entire decade to 1993 ($309bn).

Despite adding to the income of US residents, these profits do not enter into US GDP figures. Instead, they are counted in the location where they are registered – artificially inflating these countries’ GDP figures. This might be a developing country whose workers produced the profit, but saw nothing of it in the way of income. It might even be a tax haven, with no connection to the production process whatsoever.

It isn’t just investors taking advantage of looser capital controls, however. The growth of migrant labour has led to a surge in remittances, with expatriate workers sending part of their earnings back home. Despite being an essential component of household incomes in many low-income countries, remittances are entirely excluded from GDP.

The difference between GNP and GDP varies over time and between countries, meaning that the omission of capital flows from GDP has serious consequences for comparisons over time and space.

There is a more fundamental problem with using any single statistic as a measure of general prosperity, however – inequality. By compressing a national economy into a single figure, neither GDP nor GNP can tell us anything about the distribution of income.

For instance, the former Soviet republics Belarus, Georgia, Latvia and Kyrgyzstan all saw GDP per capita growth upwards of 8% in 1997, even as their median incomes plummeted – by 66% in the case of Belarus. Similarly, UK GDP per capita grew by 9.7% between 2009 and 2012, yet median incomes fell by 1.3% over the same period.

Will GDP ever be dislodged?

The problems with GDP as a general measure of social progress have been well known for decades, and alternative measures have proliferated. A decade ago, the total number of alternative indices was estimated at about 900. Today, the number may be well over 1,000.

There are several underlying approaches. One approach, known as ‘green accounting’, is to take GDP and subtract the estimated value of resource depletion and environmental degradation. Another approach bypasses market measures entirely in favour of survey-based subjective estimates of well-being, such as Bhutan’s measure of Gross National Happiness.

A further set of approaches make use of a broad array of statistics, including health, education and GDP, either aggregating these into a single index (such as the UN’s Human Development Index) or keeping them separate as a ‘dashboard’ of measures (such as the UN’s Sustainable Development Goals).

So far, however, all these efforts have produced little in the way of a serious challenge to the hegemony of GDP – if only due to sheer convenience.

Policymakers looking to measure their country’s GDP can rely on an international network of statistical agencies to proffer advice and a single unifying framework, the UN’s System of National Accounts (SNA), to ensure harmonisation. They can also rest assured that their findings will be widely disseminated by international organisations, reported on in the press and broadly understood by the public.

Moreover, many countries established their statistical infrastructure with GDP in mind – to reorient it to capture entirely new economic phenomena might be beyond the strained budgets of lower-income countries.

For Dutch economist Hoekstra, one of the key problems is that these alternative indicators are not attached to the same kind of broad array of policymaking tools and forecasting models that have grown up around GDP and the other closely related statistics in the SNA.

“What is so powerful about the SNA is that it is used in all these macroeconomic models,” he says. “Reform efforts far too often assume that if all government statistical offices suddenly adopted their preferred index that the world would suddenly change.

“But macroeconomists are offering governments far more than just retrospective GDP statistics. They have computational general equilibrium models, cost-benefit analyses, projections – all based on the GDP or other SNA statistics. Meanwhile, projections of beyond-GDP indicators are very rare.”

If this sounds like a damning critique of three decades of sustainability initiatives, Hoekstra does also see reasons for optimism.

“New Zealand recently introduced a ‘well-being budget based on a set of alternative indicators,” he says. “Crucially, its Treasury has developed tools for policymaking on top of that, such as social cost-benefit analyses, which allow them to use those indicators to allocate budgets and set priorities. 

“If we want to progress towards a society based on well-being, sustainability and equity then we need to also provide tools to decision makers.”

GDP was borne out of a period of acute need – tackling the Great Depression without reliable estimates of national income has been described as akin to trying to fly an airplane without instruments. As governments wake up to the reality of the climate crisis and spiralling inequality, they may once more find themselves flying blind. 

Ben van der Merwe

Ben van der Merwe is a data journalist at Global Data Media, specialising in FDI.